Mercurial > dovecot > core-2.2
graph
-
node color needs only one bit, not a full 32bit integer. so moved it to HEADMon, 28 Oct 2002 09:51:50 +0200, by Timo Sirainen
-
changed t_push() and t_pop() to return unsigned int. added global HEADMon, 28 Oct 2002 06:50:14 +0200, by Timo Sirainen
-
fsck() should reset lowwater marks to end of uids, not to beginning. HEADMon, 28 Oct 2002 06:36:43 +0200, by Timo Sirainen
-
keep i_buffer_seek() void and make it close the buffer if any errors happen. HEADMon, 28 Oct 2002 06:33:00 +0200, by Timo Sirainen
-
I/O buffers now use real blocking instead of setting up a sub-ioloop to HEADMon, 28 Oct 2002 06:18:26 +0200, by Timo Sirainen
-
and fix linux-sendfile check so it fails with freebsd :) HEADMon, 28 Oct 2002 05:43:13 +0200, by Timo Sirainen
-
buggy checks. we were supposed to compile them, not run. plus fixes for HEADMon, 28 Oct 2002 05:42:14 +0200, by Timo Sirainen
-
Support FreeBSD-compatible sendfile(). Completely untested. HEADMon, 28 Oct 2002 05:35:05 +0200, by Timo Sirainen
-
Moved several fields from .imap.index file to .imap.index.data file. Fixed HEADSun, 27 Oct 2002 08:37:18 +0200, by Timo Sirainen
-
Don't bother going through all that trouble to send file size to ibuffer. HEADSun, 27 Oct 2002 00:11:30 +0300, by Timo Sirainen
-
Print error message if madvise() fails (it never should). HEADSat, 26 Oct 2002 23:35:56 +0300, by Timo Sirainen
-
i_buffer_set_read_limit() - don't fail if we don't know the file size. HEADSat, 26 Oct 2002 23:06:27 +0300, by Timo Sirainen
-
mbox locking fixes / changes. we now read-lock the mbox file before syncing HEADSat, 26 Oct 2002 22:38:37 +0300, by Timo Sirainen
-
If Content-Length header was invalid, we didn't fix it properly. HEADFri, 25 Oct 2002 06:29:53 +0300, by Timo Sirainen
-
Header parsing should always keep one character lookahead, so it can see if HEADFri, 25 Oct 2002 06:27:56 +0300, by Timo Sirainen
-
Set node variable after mmap updating which may change it.. HEADFri, 25 Oct 2002 06:23:36 +0300, by Timo Sirainen
-
Don't crash either if expunge or update_flags sync_callbacks haven't been set HEADFri, 25 Oct 2002 05:52:02 +0300, by Timo Sirainen
-
Small mbox tweaks, hopefully resulting in a bit better performance. HEADFri, 25 Oct 2002 05:31:19 +0300, by Timo Sirainen
-
Tree file wasn't correctly re-mmap()ed when sync_id changed. HEADFri, 25 Oct 2002 05:25:52 +0300, by Timo Sirainen
-
Error message fixes for APPEND. Make sure we don't crash if sync_callbacks HEADFri, 25 Oct 2002 04:59:36 +0300, by Timo Sirainen
-
If i_buffer_read() fails because of EOF, set buf_errno to 0. HEADFri, 25 Oct 2002 04:58:43 +0300, by Timo Sirainen
-
Buffer returns const type, don't cast it away. HEADFri, 25 Oct 2002 04:06:19 +0300, by Timo Sirainen
-
If mbox file's timestamp isn't same as index file's, it means the file has HEADFri, 25 Oct 2002 04:01:18 +0300, by Timo Sirainen
-
imap_parser_read_args() didn't correctly always return "need for more data" HEADFri, 25 Oct 2002 01:39:42 +0300, by Timo Sirainen
-
Don't try to handle CR-only line changes, makes overly complex code and HEADFri, 25 Oct 2002 00:54:34 +0300, by Timo Sirainen
-
after _mail_tree_grow(), the node_base changes and crashes HEADThu, 24 Oct 2002 05:15:11 +0300, by Timo Sirainen
-
Instead of keeping "unused nodes" list in tree file, just move the last node HEADThu, 24 Oct 2002 04:48:33 +0300, by Timo Sirainen
-
Make sure we never truncate ourself smaller than INDEX_FILE_MIN_SIZE. HEADThu, 24 Oct 2002 04:47:58 +0300, by Timo Sirainen
-
Check fdatasync() from -lrt if it's not in libc (solaris). Also fallback to HEADThu, 24 Oct 2002 03:34:20 +0300, by Timo Sirainen
-
Fixes to timezone handling which were handling quite badly. added HEADThu, 24 Oct 2002 03:15:38 +0300, by Timo Sirainen
-
Ignore and leave maildir flags after ',' character (eg. file:2,S,U123). HEADThu, 24 Oct 2002 00:07:37 +0300, by Timo Sirainen
-
file_wait_lock() now takes timeout-parameter so one badly behaving process HEADWed, 23 Oct 2002 23:41:35 +0300, by Timo Sirainen
-
Use fdatasync() instead of fsync() where possible. msync() all files first, HEADWed, 23 Oct 2002 22:49:23 +0300, by Timo Sirainen
-
if we're only process accessing modify log when it gets full, we can just HEADWed, 23 Oct 2002 20:15:36 +0300, by Timo Sirainen
-
Modify log was pretty much broken whenever it was switched. HEADWed, 23 Oct 2002 20:03:39 +0300, by Timo Sirainen
-
input buffer limit wasn't reset in error conditions. HEADWed, 23 Oct 2002 20:02:42 +0300, by Timo Sirainen
-
Assert i_snprintf() string lengths to INT_MAX instead of SSIZE_T_MAX since HEADWed, 23 Oct 2002 16:19:56 +0300, by Timo Sirainen
-
Check vsnprintf() return values more carefully, thanks to Thomas Wouters HEADWed, 23 Oct 2002 16:16:50 +0300, by Timo Sirainen
-
Still more bugfixes to o_buffer_send_ibuffer() without sendfile. HEADTue, 22 Oct 2002 01:35:02 +0300, by Timo Sirainen
-
OBuffer was closed if sendfile() wasn't supported. HEADTue, 22 Oct 2002 01:23:20 +0300, by Timo Sirainen
-
o_buffer_send_ibuffer() didn't correctly update inbuf->v_offset when not HEADTue, 22 Oct 2002 01:18:25 +0300, by Timo Sirainen
-
cleanups, also file_wait_lock() return value wasn't checked correctly HEADMon, 21 Oct 2002 18:01:56 +0300, by Timo Sirainen
-
Hardlink COPYing still copied the file name instead of generating a new one, HEADMon, 21 Oct 2002 03:24:54 +0300, by Timo Sirainen
-
Added setting mail_save_crlf. Few other settings cleanups. HEADMon, 21 Oct 2002 03:09:29 +0300, by Timo Sirainen
-
writev() return value wasn't checked properly. EAGAIN and EINTR shouldn't be HEADMon, 21 Oct 2002 01:12:53 +0300, by Timo Sirainen
-
include sys/socket.h, not network.h. Required with solaris+ipv6 where HEADSun, 20 Oct 2002 06:26:00 +0300, by Timo Sirainen
-
Fail if %ll is used and we don't have long long. HEADSun, 20 Oct 2002 06:19:10 +0300, by Timo Sirainen
-
Whops, this was missing from previous sync changes. HEADSun, 20 Oct 2002 06:16:39 +0300, by Timo Sirainen
-
Removed several useless handers of printf_string_upper_bound(), and added a HEADSun, 20 Oct 2002 06:15:16 +0300, by Timo Sirainen
-
i_buffer_skip() can't fail anymore, next read determines if it failed. HEADSun, 20 Oct 2002 05:53:01 +0300, by Timo Sirainen
-
Added mail_cache_fields and mail_never_cache_fields settings, plus settings HEADSun, 20 Oct 2002 05:35:51 +0300, by Timo Sirainen
-
cache_fields_later() doesn't need MailIndexRecord parameter. HEADSun, 20 Oct 2002 04:36:46 +0300, by Timo Sirainen
-
Don't call mail_index_update_header() if we don't want to cache anything. HEADSun, 20 Oct 2002 04:35:48 +0300, by Timo Sirainen
-
Instead of growing tree file always 64 nodes larger, make it n% larger just HEADSun, 20 Oct 2002 04:34:15 +0300, by Timo Sirainen
-
Added cache_fields_later() to MailIndex interface. That can be used to HEADSun, 20 Oct 2002 04:12:27 +0300, by Timo Sirainen
-
Indexes were never closed because refcount wasn't updated. HEADSun, 20 Oct 2002 04:11:03 +0300, by Timo Sirainen
-
If maildir files are unexpectedly deleted, don't treat it as error, and HEADSun, 20 Oct 2002 03:17:21 +0300, by Timo Sirainen
-
When doing EXPUNGE, if there's been external flag changes in messages we're HEADSun, 20 Oct 2002 02:33:56 +0300, by Timo Sirainen
-
Modify log now stores the changes in ranges, so store 1:100 doesn't HEADSun, 20 Oct 2002 02:09:56 +0300, by Timo Sirainen
-
mail_custom_flags_list_unref() isn't needed, the list returned by HEADSat, 19 Oct 2002 19:40:17 +0300, by Timo Sirainen
-
printf_string_upper_bound(): Instead of just warning we can't handle HEADSat, 19 Oct 2002 19:31:46 +0300, by Timo Sirainen
-
MailStorage interface change. We now let storage call several sync-functions HEADSat, 19 Oct 2002 17:51:59 +0300, by Timo Sirainen
-
IMAP_ARG_EOL doesn't belong to number of arguments in list. HEADFri, 18 Oct 2002 02:57:44 +0300, by Timo Sirainen
-
If data file was compressed, other processes didn't reopen it properly. Also HEADThu, 17 Oct 2002 21:11:05 +0300, by Timo Sirainen
-
when record data had to be moved to end of .data file, the old space was HEADThu, 17 Oct 2002 20:34:25 +0300, by Timo Sirainen
-
If BODYSTRUCTURE is cached, we can use it to generate BODY reply. It takes HEADThu, 17 Oct 2002 01:54:18 +0300, by Timo Sirainen
-
ImapArgList changed from linked list to array. easier to handle that way. HEADThu, 17 Oct 2002 01:50:21 +0300, by Timo Sirainen
-
SEARCH SEEN, UNSEEN, DELETED, UNDELETED, RECENT, UNRECENT optimizations. HEADWed, 16 Oct 2002 18:05:05 +0300, by Timo Sirainen
-
SEARCH SMALLER/LARGER - we may be able to handle the request using only HEADWed, 16 Oct 2002 04:37:25 +0300, by Timo Sirainen
-
Verify memory alignment when accessing index data file. HEADWed, 16 Oct 2002 04:29:42 +0300, by Timo Sirainen
-
FETCH BODY[] sets \Seen flag to messages now while fetching it, not only HEADWed, 16 Oct 2002 04:12:50 +0300, by Timo Sirainen
-
Fixed several problems with FETCH BODY[..]<limits>, especially related to HEADWed, 16 Oct 2002 02:46:42 +0300, by Timo Sirainen
-
doing LOGOUT now prints "Aborted login" to logfile instead of "Logged out" HEADWed, 16 Oct 2002 00:01:49 +0300, by Timo Sirainen
-
Initialize random generator before chrooting so it can open /dev/urandom HEADTue, 15 Oct 2002 23:53:16 +0300, by Timo Sirainen
-
removed gnutls-kludge, we require gnutls 0.5.5 or later now. HEADTue, 15 Oct 2002 00:19:21 +0300, by Timo Sirainen
-
Use the largest output buffer size when growing corked buffer. HEADMon, 14 Oct 2002 23:26:38 +0300, by Timo Sirainen
-
check return value of o_buffer_send() correctly HEADMon, 14 Oct 2002 20:48:03 +0300, by Timo Sirainen
-
o_buffer_send_ibuffer() - handle properly writing 0 bytes HEADMon, 14 Oct 2002 18:44:35 +0300, by Timo Sirainen
-
Split IOBuffer into mmaped IBuffer, file IBuffer, memory data IBuffer and HEADMon, 14 Oct 2002 02:49:11 +0300, by Timo Sirainen
-
Check return values for malloc()s and calloc()s. HEADSun, 13 Oct 2002 18:52:19 +0300, by Timo Sirainen
-
Dropped sent_time and alignment from MailIndexRecord. SEARCH can now use HEADSat, 12 Oct 2002 01:33:54 +0300, by Timo Sirainen
-
Created datastack_mempool which is used by at least a few temporary HEADThu, 10 Oct 2002 05:01:34 +0300, by Timo Sirainen
-
changed all "temporary memory pool" references to "data stack" HEADWed, 09 Oct 2002 20:49:41 +0300, by Timo Sirainen
-
s/temporary memory pool/data stack/ which is the correct name for it. HEADWed, 09 Oct 2002 20:44:04 +0300, by Timo Sirainen
-
Parse Content-Language correctly for BODYSTRUCTURE. HEADWed, 09 Oct 2002 19:45:47 +0300, by Timo Sirainen
-
mbox_skip_forward() didn't correctly check for LF+CR+LF header ending. HEADWed, 09 Oct 2002 19:23:14 +0300, by Timo Sirainen
-
FETCH BODY[HEADER.FIELDS*] is handled now better for huge headers. For HEADWed, 09 Oct 2002 19:04:39 +0300, by Timo Sirainen
-
t_try_grow() works now for "last allocated memory in current stack frame", HEADWed, 09 Oct 2002 17:28:04 +0300, by Timo Sirainen
-
whops, update tree and modifylog with correct uid instead of 0. added extra HEADWed, 09 Oct 2002 17:03:08 +0300, by Timo Sirainen
-
Fixes for first_hole_* updating. Removed last_lookup, it doesn't speed up HEADWed, 09 Oct 2002 16:56:44 +0300, by Timo Sirainen
-
mail_tree_rebuild() wasn't updated for new indexing code HEADWed, 09 Oct 2002 02:39:19 +0300, by Timo Sirainen
-
ioloop_create() takes now pool-parameter. io_buffer_create_mmaped() takes HEADWed, 09 Oct 2002 02:26:08 +0300, by Timo Sirainen
-
Send error message if close() fails in net_disconnect(). HEADWed, 09 Oct 2002 02:22:06 +0300, by Timo Sirainen
-
We didn't destroy the ssl connection after error in send. HEADWed, 09 Oct 2002 02:20:09 +0300, by Timo Sirainen
-
imap_msgcache_get_rfc822_partial() calculated missing header size by HEADTue, 08 Oct 2002 04:15:54 +0300, by Timo Sirainen
-
"fast open" with (eg. used with STATUS) worked exactly opposite way, caching HEADMon, 07 Oct 2002 18:18:49 +0300, by Timo Sirainen
-
mail_tree_lookup_uid_range() still buggy if first_uid wasn't found. HEADMon, 07 Oct 2002 16:07:18 +0300, by Timo Sirainen
-
make io_buffer_send_flush() always removes the cork, even if there's nothing HEADSun, 06 Oct 2002 21:27:57 +0300, by Timo Sirainen
-
mail_tree_lookup_uid_range() didn't return correct seq number if the first HEADSun, 06 Oct 2002 14:51:28 +0300, by Timo Sirainen
-
dotlock fixes: don't delete dotlock when we didn't create it (when removing HEADSun, 06 Oct 2002 14:35:13 +0300, by Timo Sirainen
-
kludge: localtime() must be called before chroot(), or the timezone isn't HEADSun, 06 Oct 2002 14:03:24 +0300, by Timo Sirainen
-
Don't try to deinitialize gnutls if we haven't initialized it.. HEADSun, 06 Oct 2002 13:47:12 +0300, by Timo Sirainen
-
Write Content-Length header for rewritten messages. HEADSun, 06 Oct 2002 12:34:54 +0300, by Timo Sirainen
-
message_parse_header() now calls the function with empty name/value at end HEADSun, 06 Oct 2002 12:20:20 +0300, by Timo Sirainen
-
Added io_buffer_set_start_offset() and io_buffer_ref() and replaced HEADSun, 06 Oct 2002 08:44:27 +0300, by Timo Sirainen
-
Added autoclose_fd-flag for io_buffer_create_file() and HEADSun, 06 Oct 2002 06:09:36 +0300, by Timo Sirainen
-
When updating any of the fields, keep them in the original location in HEADSun, 06 Oct 2002 05:54:35 +0300, by Timo Sirainen
-
Small optimization for mail_index_lookup() when there's only one hole in HEADSun, 06 Oct 2002 05:15:27 +0300, by Timo Sirainen
-
If the data file was fully truncated, the mmap wasn't updated and it crashed HEADSun, 06 Oct 2002 04:10:51 +0300, by Timo Sirainen
-
First version of binary tree file, still some locking issues while opening HEADSun, 06 Oct 2002 03:30:14 +0300, by Timo Sirainen
-
Define CMSG_SPACE always if it's undefined, not only if __sun__ is defined. HEADSat, 05 Oct 2002 23:06:04 +0300, by Timo Sirainen
-
message_send() didn't correctly add CR to 256k boundaries. HEADFri, 04 Oct 2002 20:15:09 +0300, by Timo Sirainen
-
() was missing around envelope in body/bodystructure. HEADTue, 01 Oct 2002 13:58:16 +0300, by Timo Sirainen
-
message_part_serialize_update_header() was still broken, it modified the HEADMon, 30 Sep 2002 23:02:11 +0300, by Timo Sirainen
-
Cached MessagePart reading wasn't working because alignment broke the strict HEADMon, 30 Sep 2002 02:08:18 +0300, by Timo Sirainen
-
s/--enable/--with/ to all auth modules. added --with-file-offset-size=BITS HEADSun, 29 Sep 2002 20:41:16 +0300, by Timo Sirainen
-
Modify log fixes. STORE and SEARCH didn't handle properly message sequence HEADSun, 29 Sep 2002 19:19:53 +0300, by Timo Sirainen
-
Send RECENT after all EXISTS replies. Check for new mail in mailbox after HEADSat, 28 Sep 2002 12:51:13 +0300, by Timo Sirainen
-
if auth processes were killed, login process didn't reconnect to them HEADFri, 27 Sep 2002 17:16:52 +0300, by Timo Sirainen
-
Always return ENVELOPE, even if full of NILs. HEADThu, 26 Sep 2002 23:26:56 +0300, by Timo Sirainen
-
message_part_serialize_update_header() didn't update positions for subparts HEADThu, 26 Sep 2002 20:06:57 +0300, by Timo Sirainen
-
When selecting a mailbox and we're out of disk space, send ALERT to user HEADThu, 26 Sep 2002 00:25:40 +0300, by Timo Sirainen
-
If authentication failed, the failure message was written to output buffer HEADWed, 25 Sep 2002 19:03:23 +0300, by Timo Sirainen
-
Rewrite only needed parts of mbox. We don't rename() anymore, which means we HEADTue, 24 Sep 2002 19:39:00 +0300, by Timo Sirainen
-
io_buffer_send_iobuffer() didn't set buf_errno if sendfile() failed. HEADTue, 24 Sep 2002 19:36:52 +0300, by Timo Sirainen
-
One digit day wasn't parsed correctly in From-line. HEADTue, 24 Sep 2002 19:17:39 +0300, by Timo Sirainen
-
We didn't allow saving messages longer than 8192 bytes. Now we also send the HEADTue, 24 Sep 2002 16:01:47 +0300, by Timo Sirainen
-
BODY / BODYSTRUCTURE generator didn't like if message didn't start from zero HEADTue, 24 Sep 2002 10:01:32 +0300, by Timo Sirainen
-
mail_index_update_headers() seeked always to beginning of file, instead of HEADTue, 24 Sep 2002 09:55:40 +0300, by Timo Sirainen
-
CREATE and DELETE fixes for mbox (last changes for .98 :) HEADMon, 23 Sep 2002 20:23:03 +0300, by Timo Sirainen
-
Actually, first don't do anything if there's no dirty flags. Then fsck(). HEADMon, 23 Sep 2002 20:02:15 +0300, by Timo Sirainen
-
fsck() before checking the rewrite-flag, so we won't write the 1-byte HEADMon, 23 Sep 2002 19:41:09 +0300, by Timo Sirainen